npinto asgd issues - Githubissues

npinto / asgd

Averaged Stochastic Gradient Descent Classifiers

41 stars 21 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Parameter for (exponentially) moving average

#23 npinto opened 12 years ago
0
Working feedback

#22 npinto opened 12 years ago
0
Early stopping

#21 jaberg opened 12 years ago
0
auto step size

#20 jaberg closed 12 years ago
2
TheanoBinaryASGD attempt 2

#19 jaberg closed 12 years ago
4
simple_blas.h missing?

#18 jaberg closed 12 years ago
2
TheanoBinaryASGD

#17 jaberg closed 12 years ago
4
new reference for readme

#16 jaberg closed 12 years ago
1
re-parameterize annealing schedule

#15 jaberg closed 12 years ago
11
Sparsity trick on weights update

#14 npinto opened 12 years ago
0
step_size0 should be cross validated on a small subset of the data

#13 npinto opened 13 years ago
3
Use η0 / (1 + λ η0 t)^0.75 by default instead of ...^2/3.

#12 npinto opened 13 years ago
0
Compute X_mean and X_std in fit().

#11 npinto opened 13 years ago
0
The Cython wrapper cannot find the symbols defined by BLAS

#10 argoncloud closed 12 years ago
3
ENH: warm-up period with exponential moving asgd and switch from sgd to asgd when empirical loss gets higher

#9 npinto opened 13 years ago
0
Online LOOE-like for hyper-parameters selection.

#8 npinto opened 13 years ago
0
Add Naive Multiclass "One-Versus-All" (OVA) ASGD with tests for both Binary and OVA.

#7 npinto closed 13 years ago
5
multiclass

#6 yamins81 closed 13 years ago
1
l2_regularization can be 0

#5 jaberg closed 13 years ago
3
Weight updates only if margin constraint violated.

#4 npinto opened 13 years ago
1
kwargs for mini_batch settings

#3 npinto opened 13 years ago
0
Ready to test!

#2 argoncloud closed 13 years ago
5
BLAS and heap safety

#1 argoncloud closed 13 years ago
3